struct nf_sockopt_ops

struct nf_sockopt_ops

[Carl-Daniel Hailfinger]

Forgive me if that has already been discussed.

I'm wondering why the get and set members of the struct take
some form of int as last parameter. Wouldn't be a long be the
more logical choice? I can see changing the type here wouldn't
help because nf_sockopt also accepting only an int for len.

struct nf_sockopt_ops
{
/* snip */
        int (*set)(struct sock *sk, int optval, void __user *user, unsigned int len);
/* snip */
        int (*get)(struct sock *sk, int optval, void __user *user, int *len);
/* snip */
};

Or at least the last parameter of the set member being a
signed int to match the parameter to nf_sockopt.

Any insight into why the parameters were chosen that way
would be truly appreciated.


Regards,
Carl-Daniel

Re: struct nf_sockopt_ops

[Patrick McHardy]

Carl-Daniel Hailfinger wrote:
> Forgive me if that has already been discussed.
> 
> I'm wondering why the get and set members of the struct take
> some form of int as last parameter. Wouldn't be a long be the
> more logical choice? I can see changing the type here wouldn't
> help because nf_sockopt also accepting only an int for len.
> 
> struct nf_sockopt_ops
> {
> /* snip */
>         int (*set)(struct sock *sk, int optval, void __user *user, unsigned int len);
> /* snip */
>         int (*get)(struct sock *sk, int optval, void __user *user, int *len);
> /* snip */
> };
> 
> Or at least the last parameter of the set member being a
> signed int to match the parameter to nf_sockopt.
> 
> Any insight into why the parameters were chosen that way
> would be truly appreciated.

For compatibility. setsockopt(2):

The fifth argument of getsockopt and setsockopt is in  reality  an  int
[*]  (and  this  is what BSD 4.* and libc4 and libc5 have).  Some POSIX
confusion resulted in the present socklen_t, also used by  glibc.   See
also accept(2).

Regards
Patrick